Alcoholic beverages are well-established as risk factors for cancer at several sites, but it is not certain to what extent the increased risk is attributable directly to the ethanol content of these beverages, or to possible contaminants or congeners. We propose an epidemiologic study to evaluate the relative carcinogenicity of various alcoholic beverages, comparing clear liquors (vodka, gin, light rum), which have few contaminants or congeners, with smoky liquors, beer, red wine and white wine. The study design will be a population-based case-control study, including patients with laryngeal and hypopharyngeal cancer, and population controls. Case ascertainment will be arranged directly with otolaryngologists in the Boston area. Controls will be selected in a ratio of three controls per case; we shall accrue subjects for 36 months, during which we hope to interview 3000 subjects, of whom 750 will be patients with laryngeal or hypopharyngeal cancer.